Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Zhang, Yijia, Gou, Zhihong, Cao, Shijie, Feng, Weigang, Zhang, Sicheng, Dai, Guohao, Xu, Ningyi
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2411.18873
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866912136365604864
author Zhang, Yijia
Gou, Zhihong
Cao, Shijie
Feng, Weigang
Zhang, Sicheng
Dai, Guohao
Xu, Ningyi
author_facet Zhang, Yijia
Gou, Zhihong
Cao, Shijie
Feng, Weigang
Zhang, Sicheng
Dai, Guohao
Xu, Ningyi
contents Deep Neural Networks (DNNs) have revolutionized various fields, but their deployment on GPUs often leads to significant energy consumption. Unlike existing methods for reducing GPU energy consumption, which are either hardware-inflexible or limited by workload constraints, this paper addresses the problem at the GPU kernel level. We propose a novel search-based compilation method to generate energy-efficient GPU kernels by incorporating energy efficiency into the search process. To accelerate the energy evaluation process, we develop an accurate energy cost model based on high-level kernel features. Furthermore, we introduce a dynamic updating strategy for the energy cost model, reducing the need for on-device energy measurements and accelerating the search process. Our evaluation demonstrates that the proposed approach can generate GPU kernels with up to 21.69% reduced energy consumption while maintaining low latency.
format Preprint
id arxiv_https___arxiv_org_abs_2411_18873
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Automating Energy-Efficient GPU Kernel Generation: A Fast Search-Based Compilation Approach
Zhang, Yijia
Gou, Zhihong
Cao, Shijie
Feng, Weigang
Zhang, Sicheng
Dai, Guohao
Xu, Ningyi
Performance
Machine Learning
Deep Neural Networks (DNNs) have revolutionized various fields, but their deployment on GPUs often leads to significant energy consumption. Unlike existing methods for reducing GPU energy consumption, which are either hardware-inflexible or limited by workload constraints, this paper addresses the problem at the GPU kernel level. We propose a novel search-based compilation method to generate energy-efficient GPU kernels by incorporating energy efficiency into the search process. To accelerate the energy evaluation process, we develop an accurate energy cost model based on high-level kernel features. Furthermore, we introduce a dynamic updating strategy for the energy cost model, reducing the need for on-device energy measurements and accelerating the search process. Our evaluation demonstrates that the proposed approach can generate GPU kernels with up to 21.69% reduced energy consumption while maintaining low latency.
title Automating Energy-Efficient GPU Kernel Generation: A Fast Search-Based Compilation Approach
topic Performance
Machine Learning
url https://arxiv.org/abs/2411.18873